Hi Victor,
Keep looking through the mailing list archive for more things to try.
I've had success with some of the following:
- Reduce the timestep (dt)
- Try using noshakemask on the entire solute.
- If applicable, make sure ig is set to the same number for v0 and
v1 (see amber11 bugfix 13)
--Niel
________________________________________
From: Victor Ma [victordsmagift.gmail.com]
Sent: Friday, August 26, 2011 10:19 AM
To: AMBER Mailing List
Subject: [AMBER] problem with TI calculation, help!!
hi all
I'm working on the thermodynamic integration calculation. I'm trying to
change a CH3 group to CF3 group. I initially did this directly and the
simulation did not converge very well. So I thought changing three heavy
atoms at the same time might be too much. I decided to do it one by one.
First I'd like to change CH3 to CH2F. However I'm seeing errors like this:
vlimit exceeded for step 62685; vmax = 37.6809
Coordinate resetting (SHAKE) cannot be accomplished,
deviation is too large
NITER, NIT, LL, I and J are : 0 2 2 63 67
These are the distances between C and H/F in the CH2F group. I did some
research and tried a few things, like optimizing the initial structure,
longer minimization, slowly raising temperature,etc. So far, none of them
worked. I'm also seeing similar problem when I'm changing phenyl to phenol
(H to OH). So I'm wondering if this is because the relative big
reorganization of charge distribution between the reference compound and new
compound and the bond distance and angles between the C-H and C-F are too
different.(?)
This error occurs typically at the middle step, when changing H to F using
soft core potential. In addition, in the amber tutorial, the minimization
for this step uses SHAKE (ntf=2, ntc=2). However it never went through in my
calculation. THe program complains that you can only do steepest decent and
"softcore potentials require ntf=1 because SHAKE constraints on some bonds
might be removed". So i'm using "ntmin=2, ntf=1,ntc=1" instead. I'm not sure
if this is causing problem.
Any idea? I really appreciate it. Thanks a lot.
Victor.
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Aug 26 2011 - 10:00:04 PDT
